Selective Co‐Encapsulation Inside an M6L4 Cage

Abstract There is broad interest in molecular encapsulation as such systems can be utilized to stabilize guests, facilitate reactions inside a cavity, or give rise to energy‐transfer processes in a confined space. Detailed understanding of encapsulation events is required to facilitate functional molecular encapsulation. In this contribution, it is demonstrated that Ir and Rh‐Cp‐type metal complexes can be encapsulated inside a self‐assembled M6L4 metallocage only in the presence of an aromatic compound as a second guest. The individual guests are not encapsulated, suggesting that only the pair of guests can fill the void of the cage. Hence, selective co‐encapsulation is observed. This principle is demonstrated by co‐encapsulation of a variety of combinations of metal complexes and aromatic guests, leading to several ternary complexes. These experiments demonstrate that the efficiency of formation of the ternary complexes depends on the individual components. Moreover, selective exchange of the components is possible, leading to formation of the most favorable complex. Besides the obvious size effect, a charge‐transfer interaction may also contribute to this effect. Charge‐transfer bands are clearly observed by UV/Vis spectrophotometry. A change in the oxidation potential of the encapsulated electron donor also leads to a shift in the charge‐transfer energy bands. As expected, metal complexes with a higher oxidation potential give rise to a higher charge‐transfer energy and a larger hypsochromic shift in the UV/Vis spectrum. These subtle energy differences may potentially be used to control the binding and reactivity of the complexes bound in a confined space.


Introduction
Supramolecular chemistry has evolvedt oastage whereby al arge varietyo fw ell-defined structures have been prepared by assemblyo fs mall buildings blocks. Now that control over the shapea nd structure of these nano-sized objectsh as been well established, the introductiona nd evaluationo ff unctional properties is am ore pressing concern.S elf-assembled cages typicallyf unctiona sh osts for small molecules that can bind in their interior.T he bindingo fs uch guest molecules is generally based on ac ombination of noncovalent interactions, including hydrogen bonds,i onic or p-p interactions, and hydrophobic effects. Shape complementarity is obviously important, andi t has been established that the ideal size of the guest is about 55 %o ft he size of the interior of the cage. [1] Typically,t hese self-assembledhostsa re interestingo bjectsf or exploring chemistry with the compounds in their confined spaces. [2][3][4] Indeed, examples of capsule-induced selectivity in organic reactionsh ave been reported. [5][6][7] Furthermore, when metal catalysts are captured inside ac apsule, the cages resemble to some extent the second coordination spheret ypical of enzymes, and as such new reactivity and selectivity may be displayedb yt he encapsulated catalysts. [8][9][10][11] Also in the photosynthetic apparatus in nature,t he second coordination sphere is of importance, and in analogyc hromophoric guests have been encapsulated and subjected to detaileds tudies. As ar esult, inducedc harge-transfer (CT) complexeso re xciplexes have been enforced by the closep roximity of guests and hosts. While charge-transfer complexes can be considered as involving only aw eaki nteraction between the donor and acceptor, they can often be easily observed and characterized due their intense colors. It hasb een demonstrated that selectivee ncapsulation of guests in preformed hosts, such as cyclodextrins, [12] cucurbit [8]urils, [13,14] pillar[5]arenes, [15] porousc oordination polymers, [16,17] and metallocages, [18] can be used to obtain new charge-transfer complexes. Additionally,t he formation of charge-transfer complexes can be used as ad riving force to form new supramolecular assemblies. [19,20] Whereas the encapsulation of as ingle guest, such as an organico rm etal complex, has been reportedf requently,t he selective coencapsulation of two different guests is not straightforward. Nevertheless, this type of assembly is of particular interest as it facilitatest he occurrence of chemistry between two components, ap rerequisite for metal-catalyzed transformations. Herein, we report the selectivee ncapsulationo ft wo different guests inside an octahedral metallocage to form ternary complexes.T he ternary complex consists of aC p-ligated metal complex that is in close proximity to af lat aromatic guest, which form ac harge-transfer complex as indicated by UV/Vis spectrophotometry.D ifferent aromatic guests and metal complexes can be used to selectively form the ternary complexes, and the binding efficiency depends on the properties of the components. The resulting charge-transfer complex varies according to the oxidation potentialo ft he electron-donor metal complex.U nderstanding selective co-encapsulationo fametal complex and substrate in ac onfined space is importantf or future studies on catalysis in such environments.

Results and Discussion
Selectiveco-encapsulation and characterization of the ternary complexes As elf-assembled metallocage, which has been reported previously ( Figure 1), was used as the host for the current studies. This metallocageh as recently been demonstrated to facilitate exciplex formation with different guests [21][22][23] and to incorporate metal complexes, whereupon ac harge-transfer band was observed. [24] The water-soluble octahedral nanosphere 1 ( Figure 1) can bind al arge variety of guests,m ostly on the basis of hydrophobic effectsa nd interactions with the electron-deficient sidewalls of its cavity.Abroad array of guests have been accommodated in this molecular container,i ncluding metal complexes, [24][25][26] substrates for organic reactions, [27,28] halogens, [29,30] and av ariety of molecules with interesting UV/Vis spectrophotometric properties. [21][22][23]31] Generally,t he host-guest complexes are formed by simply mixingt he components.
We investigated whether [(CpMe)Ir(cod)] 2 (Cp = cyclopentadiene,c od = 1,5-cyclooctadiene) could be encapsulated. Stirring as uspensiono f2 and 1a in D 2 Oa t1 00 8Cf or prolonged reactiont imes did not result in significant encapsulationo f2 as no new signals were observed in the NMR spectra. Cage 1 has been reportedt ob ind some guests in the presenceo f an appropriate second molecule that is co-encapsulated. [32][33][34] In this context, we investigated binding of the metal complex in the presence of triphenylene (6)a saco-guest.H eating a suspension of 1a, 2,a nd 6 in D 2 Oa t1 00 8Cf or 1hresulted in ac olored suspension.I nterestingly,a fter cooling the solution and filtration of the excessg uests, ap urple solutionw as obtained indicating that the ternary complex 1a·2·6 had been formed (Scheme1).
Due to their low solubility in water,n either the iridium complex nor the triphenylenec ould be detected by 1 HNMR spectroscopyi nD 2 Os olution. In contrast, in the presence of the cage, their signals were clearly visible. Binding was confirmed by the clear upfield shifts displayed by the guest molecules in the 1 HNMR spectrum, as depicted in Figure 2. The shielding causedb yt he aromatic rings of the cage resulted in typical shifts of the triphenylene signals of 1.7 and 1.4 ppm, whereas the signals of the metal complex were shifted by between 2.5 and 3.2 ppm. Another indication of guest encapsulationw as provided by the change of symmetry [34] of the metallocage from T d to effectively C 3 (on the NMR timescale), as was clear from the 1 HNMR spectra. This loss of symmetry occurs because each guest occupies one half of the capsule. The guests are closely packed against the triazine panel and on the NMR timescale the cage loses its T d symmetry [35] (see Supporting Information S4). For the empty cage, all pyridine rings are equivalent so that only two signals are observed for the pyridine protons. However,f or the cage incorporatingt he two guests, eight sets of pyridine protons ignals are observed. Importantly, the symmetry of the guest molecules was not affectedb yt he encapsulation, indicating that these can freely rotate inside the void of the capsule. Additional support for guest encapsulation was provided by diffusion-ordered NMR (DOSY,d epicted in Supporting Information S8), which showed that the diffusion constant of the guests matched that of the capsule. [24] Due to the enforced close proximity of the respectiveg uests, NOE Scheme1.Co-encapsulation of guests 2 and 6 in host 1a. Attempts to crystallize the ternary product using metallocage 1a didn ot yield crystalline materials uitable for X-ray diffraction analysis. However,w hen we changed the nanosphere to its palladium analoguew ith a2 ,2'-bipyridine cis-capping ligand (metallocage 1b), [36] crystals suitablef or X-ray diffraction analysisw ere obtained. The data revealed am olecular structure in which the two guests occupy the void of the sphere (Figure 3), showing at ight fit of both guests inside the cage. The guest pair (2·6)w as disordered over three positions, hence only the 33.3 %o ccupancy is displayed to clearly show the host-guest structure. Although the solid-state structure cannot be directly compared with the structure in solution,t he disordero vert hree positions correspondst ot he effective C 3 symmetry of the host-guest assembly on the NMR timescale, as derived from the NMR spectra in solution at room temperature.
The molecular structure shows that the triphenylene guest (6)r esides close to the triazine panel. However,i ti sn ot stacked completely parallel to the panel.T he distance from 6 (central ring) to the central ring of the triazine panel is 3.48 ,  and that from 6 to the iridium atomi s4 .84 .T he average angle between the pyridine moieties of the triazine panels that are connected to the palladium centeri s8 6.48.T his is similar to the average of 86.68 previously reportedf or the crystal structure of the cage incorporatinga na damantoid water molecule cluster. [36b] Scope of the co-encapsulation Following the observation that only the two guests together could be embedded inside the cavity of the metallocage and that no individual encapsulation occurs, the scope of the metal complexes and aromatic moieties that can be co-encapsulated was explored. The 1 HNMR spectrum of 1c·2·6 showedi ncomplete encapsulation as empty cage 1c was still present. Integration of the pyridinep rotons ignals and comparison of these with the aromatic signals allowed determination of the efficiency of the formation of the ternary complex, expressedi n terms of percentageo fo ccupied cage;t his was 75 %f or 1c·2·6.N ext, the metal complex was changed to the sterically similar,b ut electronically distinct, rhodiumc omplex 4,w hich resultedi nasimilar amount of ternary complex (see Table 1).
Interested by this difference in the formation of the co-encapsulated species, we changed the aromatic guest to pyrene (7)o rp erylene ( 8). The ternary complexes formed with pyrene showedahigher amounto fc o-encapsulation. With guest 7, high formation of ternary complex was also observed with the sterically hindered [(CpMe 4 )Rh(cod)] (81 %). Conversely,c o-encapsulation with perylener esulted in lower amountso ff ormation of the ternary complexes, among which that of 1c·5·8 was the lowest, amounting to am ere 14 %. The more efficient formation of ternary complexes with pyrene compared to triphenylene prompted us investigate whether the former guest could displace the latter.As olution of the triphenylene ternary complex 1c·3·6 was therefore mixed with 10 equivalents of pyrene at 100 8Cf or 1h (depicted in Scheme2,t op). After cooling and filtration of the excess guest, it was clear that exchange of the aromatic guest had taken place, and the ratio of 1c·3·6 to 1c·3·7 was determined as 1:6.3. Similarly,asolution of 1c·3·7 was mixed with triphenylene( 6)a t1 00 8Cf or 1h.I n this case, the ratio of 1c·3·6 to 1c·3·7 was 1:36.0. Thus,p yrene easily displaces triphenylene and has ah igher affinity for the cavity of the metallocage.
To furtherp rove that perylenei st hermodynamically the least favored in the void of metallocage 1,amixture of excesses of all three aromatic guests (20 equivalents) was stirred with 3 and 1c.A fter 1h at 100 8C, this resulted in ar atio of 1c·3·6:1c·3·7 of 1:9.1 and no perylenewas encapsulated. Heating the suspension with the guestsf or al onger time (6 h) slightly influenced the ratio (1:10.3) buts till did not result in the encapsulationof8.This demonstrates that the metallocage has different affinitiesf or the aromatic guests and that this is represented in the amount of the ternary complex that is formed.

Tuning of the charge-transfer band through co-encapsulation
Intrigued by the distinct colors that were observed with the different ternary complexes,w ep roceeded to examine the coencapsulated species by UV/Vis spectrophotometry.U pon coencapsulation of rhodiumc omplex 4 and triphenylene,t he pale-yellow solution of empty cage 1c turned purple and an ew absorption band was observed at 555 nm (2.23 eV). This new band is indicative of the formation of ac harge-transfer complex (see Figure4). To confirm that the cage elicits the new absorption band, we combined the compoundsi nc hloroform. The absence of ac harge-transfer band in the UV/Vis spectrum of this solutiond emonstrated that pre-organization in the metallocage is required for exciplex formation.
The formation of charge-transferc omplexes of soluble molecules in supramolecular complexes based on cucurbit [8]uril Scheme2.Exchange studies of preformed ternary complexeswith pyrene (top)o rt riphenylene (bottom). These experiments showed the metallocage to haveapreference for binding pyrene. Table 1. Formation of the ternary complexes in metallocage 1c with metal donors 2-5 and aromatic guests 6-8.A ll encapsulation studies were performedw ith the samen umberso fe quivalents of guests for 1h at 100 8C. Percentages are based on integrationo ft he pyridine 1 HNMR signals.

Co-encapsulation [%]
Triphenylene (6)P yrene (7) has been reported previously. [13,15,37] In the current system, formation of the host-guest system is based on extraction of the guests,w hicha re essentially insoluble in water.T he metal complexes that are co-encapsulated by this metallocage make it possible to investigate the effect of the redox potential on the charge-transfer complex. In this regard, various pairs of guests for the formation of ternary complexes were investigated. The donori nt he charge-transfer complex was expected to be the metal complex, hence various complexes with different steric and electronic properties weres tudied. The same steric structurew as retained and only the electronic properties were changed by utilizing[ (CpMe)Ir(cod)] (2)f or the encapsulation studies. When 2 was co-encapsulated with triphenylene, the charge-transfer band was shifted to higher energy (513 nm, 2.42 eV;s ee Ta ble2)c ompared to that obtained with the rho-dium complex (4). This blue-shift of the charge-transfer band was anticipated as the oxidation potentialo f5 is higher than that of 3,a si ndicated by the redox potentials of the complexes in dichloromethane determined by cyclic voltammetry measurements (see Ta ble2 and SupportingI nformation S6). When the electronic properties of the rhodiumc omplex were furtherc hanged by substituting it with four methyl groups (complex 5)a sopposed to no methyl groups (complex 3), it becamea pparent that metal complexes with ah igher oxidation potential yieldedacharge-transfer band with triphenylene at lower wavelength.T ot he best of our knowledge, this is the first time that metal complexes have been used to fine-tune charge-transfer bands through the formation of ternary complexes, based on their redox potentials. To ascertainw hether similar trends could be observed, we co-encapsulated other aromatic guests. For this purpose, pyrene (7)a nd perylene ( 8) were used to form the ternary complexes. With perylene( 8)a s aromatic co-guest, the charge-transfer bands shifted to higher energies. Ar ecurringt rend in the energies was observed, with the exceptiono f1c·2·8,f or which the energy was difficult to determine due to overlap with the UV band of perylene. The use of pyrene resulted in almostq uantitative formation of the ternary complexes (see Ta ble 1), including with the sterically demanding 5.H owever,t he locationo ft he charge-transfer band was difficult to determine precisely as it partially overlapped with the absorption of pyrene (see Supporting Information S5). Although the estimated energies were therefore less accurate, all CT bands shiftedt ol ower energies as ac onsequenceoft he lower-lying LUMO of pyrene. This was consistent with the metal complex being the electron donor and the aromatic compound the acceptor.T he resulting charge-transfer complexm ay be stabilized by the electron-poorc avity of the molecular container,b ut the dominant effect is the chargetransfer complexf ormation between the two guests as variation of either component leads to changes in the bands. Based   [b] 2.50 [(CpMe)Ir(cod)] (2)0.122 .42 2.04 [b] 2.38 [b] [a] Oxidationp otentials were measuredu sing 1mm solutions of the metalc omplex in CH 2 Cl 2 containing 0.1 m TBAPF 6 at aglassy carbon working electrode. The potentials are referenced to ferrocene (Fc 0/ + )a nd based on simultaneous fitting of an irreversible wave at multiple scan rates (0.1/0.3/1.0Vs À1 ); see Experimental SectionS 6f or mored etails. [b] Due to overlap of the CT bandw ith the UV band of the aromatic compound, the exact CT energy is difficult to determine and less accurate.
on thesef indings, we constructed as chematic energy diagram for the charge-transfer interactions, asd epicted in Figure 5.

Conclusion
We have demonstrateds elective co-encapsulation in metallocage 1,t he void of which is occupied by one metal complex and one flat aromatic guest. As the individual guests are not encapsulated under the applied conditions, this resultr epresents an example of selectivec o-encapsulation. The shapes of the respective guests complement each other ando nly together fill the void of the cage. The amount of co-encapsulation depends on the guests, and varies with their steric properties. Exchange of the guests is possible, and offering multiple suitable guests shows at hermodynamic preference for certain ternary complexes.I nt he UV/Vis spectra of these complexes, clear charge-transfer bands are observed, the energy of which is controlled by the electronic properties of the donor (metal complex) and the acceptor( aromatic guest). The void of the metallocage makesi tp ossible to study the formation of varioust ernary complexes,s howing the benefits of these host-guest systems. The possibility of selectively co-encapsulating two guests is important in the future design of novel host-guest complexes,a nd could ultimately lead to catalytic transformations insidethese metallocages.